Epilepsy Research (2014) 108, 782—791
journal homepage: www.elsevier.com/locate/epilepsyres
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Ryan G. Wagner a,b,c,∗, Anthony K. Ngugi a,d,e, Rhian Twine a, Christian Bottomley f,g, Gathoni Kamuyu d, F. Xavier Gómez-Olivé b, Myles D. Connor b,h, Mark A. Collinson b,c, Kathleen Kahn a,b,c, Stephen Tollman a,b,c, Charles R. Newton a,b,d,i,j,k a
Studies of Epidemiology of Epilepsy in Demographic Surveillance Systems (SEEDS)—INDEPTH Network, Accra, Ghana b MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa c Division of Epidemiology and Global Health, Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden d KEMRI/Wellcome Trust Research Programme, Centre for Geographic Medicine Research—Coast, Kilifi, Kenya e Research Support Unit, Faculty of Health Sciences, Aga Khan University- East Africa, Nairobi, Kenya f Department of Infectious Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom g MRC Tropical Epidemiology Group, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom h Borders General Hospital, Melrose, United Kingdom i Neurosciences Unit, UCL Institute of Child Health, London, United Kingdom j Clinical Research Unit, London School of Hygiene and Tropical Medicine, London, United Kingdom k Department of Psychiatry, University of Oxford, Oxford, United Kingdom Received 26 May 2013; received in revised form 21 November 2013; accepted 14 January 2014 Available online 29 January 2014
KEYWORDS Epilepsy; Prevalence;
Summary Rationale: Epilepsy is among the most common neurological disorders worldwide. However, there are few large, population-based studies of the prevalence and risk factors for epilepsy in southern Africa.
∗ Corresponding author at: MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), PO Box 2 Acornhoek 1360 South Africa. Tel.: +27 13 795 5076; fax: +27 13 795 5076. E-mail address: [email protected] (R.G. Wagner).
0920-1211/$ — see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2014.01.004
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Case-control; Risk factors; Population-based
783
Methods: From August 2008 to February 2009, as part of a multi-site study, we undertook a threestage, population-based study, embedded within the Agincourt health and socio-demographic surveillance system, to estimate the prevalence and identify risk factors of active convulsive epilepsy (ACE) in a rural South African population. Results: The crude prevalence of ACE, after adjusting for non-response and the sensitivity of the screening method, was 7.0/1,000 individuals (95%CI 6.4—7.6) with significant geographic heterogeneity across the study area. Being male (OR = 2.3; 95%CI 1.6—3.2), family history of seizures (OR = 4.0; 95%CI 2.0—8.1), a sibling with seizures (OR = 7.0; 95%CI 1.6—31.7), problems after delivery (OR = 5.9; 95%CI 1.2—24.6), and history of snoring (OR = 6.5; 95%CI 4.5—9.5) were significantly associated with ACE. For children, their mother’s exposure to some formal schooling was protective (OR = 0.30; 95%CI 0.11—0.84) after controlling for age and sex. Human immunodeficiency virus was not found to be associated with ACE. Conclusions: ACE is less frequent in this part of rural South Africa than other parts of sub-Saharan Africa. Improving obstetric services could prevent epilepsy. The relationship between snoring and ACE requires further investigation, as does the relative contribution of genetic and environmental factors to examine the increased risk in those with a family history of epilepsy. © 2014 Elsevier B.V. All rights reserved.
Introduction Epilepsy is one of the most common neurological disorders in the world, affecting about 69 million people worldwide, with 90 percent living in low- and middle-income countries (LMICs) (Ngugi et al., 2010). It contributes nearly one percent to the global burden of disease (Murray et al., 2012), and 20 percent of the global burden of epilepsy is in Africa (World Health Organization, 2004). While these figures suggest a large burden of epilepsy in Africa, they are derived from a limited number of studies that employ different case definitions and methodologies. Studies suggest that utilizing hospital records in LMICs to detect epilepsy under-estimates the prevalence by at least 80 percent due to limited health care utilization by people with epilepsy (PWE) in these settings (Osuntokun et al., 1987). Thus, population-based surveys are frequently used to estimate prevalence, though not without limitations, including the absence of well-demarcated populations and vital statistics registries. This limitation, coupled with the lack of trained medical personnel available to make the diagnosis of epilepsy, makes estimating the burden of epilepsy in sub-Saharan Africa unusually challenging. A recent systematic review and meta-analysis highlighted the significant variation in the prevalence of epilepsy between high-income countries and LMICs, with a higher prevalence in LMICs, especially in rural settings (Ngugi et al., 2010). The authors suggest that study size and the economic development level of the study country largely explain the heterogeneity, although increases in obstetric injury, head injuries, and infections and infestations of the central nervous system (CNS) (Newton & Garcia, 2012), such as toxoplasmosis and toxocara (Wagner & Newton, 2009), malaria (Carter et al., 2004) and human immunodeficiency virus (HIV) are thought to contribute (Ngugi et al., 2010), but there is little data from South Africa where the prevalence of HIV is very high. As part of a multi-centre study on the epidemiology of epilepsy in demographic sites (SEEDS) (Ngugi et al., 2013), we conducted a three-stage, population-based survey and a case-control study to determine the prevalence of and risk factors for active convulsive epilepsy (ACE) in rural South
Africa. In particular, we were interested in examining the risk factors for ACE in a non-malaria endemic area, particularly HIV since it has a high prevalence in South Africa.
Methodology Population and study area The study was conducted in the rural Agincourt health sub-district, in which the Agincourt health and sociodemographic surveillance system (HDSS) operates and is located 500 kilometers northeast of Johannesburg (Fig. 1). The Agincourt HDSS was established in 1992 as a research platform to inform health and development policy through evidence-based research (Kahn et al., 2012). The population has been enumerated through an annual census update, following baseline measurement in 1992 and captures vital statistics including births, deaths, and in- and out-migrations. In 2008, the population was 83,121 individuals in 15,841 households and 25 villages on 420 km2 of semi-arid scrubland. The site forms part of a former bantustan, or ethnic ‘homeland’ implemented during the Apartheid era. The population is mainly Xi-Tsonga speaking, with one-third (32.8%) originally from Mozambique. The leading causes of death ascertained through the Agincourt HDSS are HIV/AIDS, cardiovascular disease and trauma (road traffic accidents, assaults) (Tollman et al., 2008). Six government clinics, one larger government health centre, and one public-private community health centre, with its main focus being HIV and tuberculosis, provide primary health care services for the population. Referrals are to three government district hospitals located 25 to 55 km from the sub-district.
Ethics Ethical clearance for the study was received from the Human Research Ethics Committee of the University of the Witwatersrand, Johannesburg, South Africa (Clearance number: M080455) and the Mpumalanga Province Department of
784
R.G. Wagner et al.
Figure 1 Map of South Africa, Bushbuckridge Municipality, and the Agincourt sub-district and health and socio-demographic surveillance system, 2008.
Health’s Research and Ethics committee. Written informed consent was sought from each participant in the study. Parental or guardian informed consent was sought in the case of children or patients with cognitive impairment, with verbal assent being sought from children and cognitively impaired patients.
Case definition of active convulsive epilepsy (ACE) We identified patients with Active convulsive epilepsy (ACE) since convulsive epilepsies are associated with increased morbidity, mortality (Diop et al., 2005) and greater social stigma than non-convulsive epilepsies (World Health Organization, 2005). ACE was defined as having two or more unprovoked convulsive seizures occurring more than 24 hours apart, with at least one seizure occurring in the 12 months preceding the study or currently taking antiepileptic drugs (AEDs) (Edwards et al., 2008).
and study neurologist (CRN) on the administration of this questionnaire. Individuals who had experienced at least two seizures in their lifetime and at least one seizure with abnormal movement (aimed at identifying convulsive epilepsy) in the preceding 12 months or currently taking AEDs were invited for further assessment in the epilepsy clinic within the following week (Stage III). In Stage III (September 2008—February 2009), sociodemographic information was obtained and a specially trained nurse made the diagnosis of ACE, based on a positive clinical history provided by the patient’s caregiver or family member. A clinical examination was completed to identify neurological abnormalities and physical co-morbidities. Blood was drawn from each patient for serological testing. A neurologist (CRN) confirmed the diagnosis of epilepsy by reviewing all of the patient records and examining selected patients.
Population sample Procedures Between August and November 2008, two questions (‘Do you have fits or has someone ever told you that you have fits?’ and ‘Do you experience episodes in which you legs or arms have jerking movements or fall to the ground and lose consciousness?’), used previously in a similar study (Edwards et al., 2008), were administered by census fieldworkers to a senior member of each household (n = 15,841) on behalf of each member of the household. The questions had been piloted previously in the local language (Xi-Tsonga) at a district hospital and had a high specificity (100%) and sensitivity (98%). The questions were included in the 2008 annual Agincourt HDSS census update round. The two questions within the census update constituted Stage I of the study and sought to identify all people who had experienced a seizure prior to 1 August 2008 (the prevalence day). In Stage II (August—December 2008), trained fieldworkers visited all individuals who responded ‘yes’ to either of the two questions in Stage I and asked a detailed questionnaire, based on previous studies (Placencia et al., 1992), about the characteristics and history of the seizures. Fieldworkers were specifically trained by the study managers (RGW, AKN)
In addition to identifying cases of epilepsy through the population screen (Stage I), a random sample of 4,500 individuals (all ages and both sexes), from the Agincourt HDSS, were screened with the Stage II tool. This population sample simulated the traditional two-stage epilepsy prevalence study design (Placencia et al., 1992) and was used to evaluate the three-stage methodology of this study. Those who screened positive were referred for assessment in Stage III and included as cases in the case-control study. Furthermore, individuals who were referred to the epilepsy clinic by sources (clinic staff and community leaders) other than the three-stage study were clinically assessed and those found to have ACE were included as cases in the case-control study.
Selection of controls for case control study Controls were selected randomly from those that screened negative for epilepsy in the population sample and frequency matched to cases diagnosed with ACE during Stage III of the study (on age-bands 0—5, 6—12, 13—18, 19—28, 29—49, and 50+ years). If the selected control was not found
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Table 1
785
Crude and adjusted prevalence of ACE by age and sex, Agincourt sub-district, South Africa 2008.
Age band
Cases of ACE
Crude prevalence
95% Confidence interval
Adjusted prevalence
95% Confidence interval
0 to 5 years 6 to 12 years 13 to 18 years 19 to 28 years 29 to 49 years 50+ years Male Female
10 29 34 47 87 38 129 116
0.87 2.23 2.76 2.65 4.70 3.94 3.24 2.69
0.33—1.41 1.42—3.03 1.83—3.68 1.89—3.40 3.71—5.69 2.68—5.19 2.68—3.80 2.20—3.18
2.18 4.79 6.38 6.23 10.95 9.55 7.43 6.50
0.88—3.48 3.09—6.52 4.32—8.42 4.47—7.98 8.74—13.13 6.69—12.41 5.53—7.82 4.53—6.54
after three attempted visits, another control was selected randomly as a replacement. For each control selected, a clinical history and blood sample were taken and a clinical examination performed.
variables on whether alcohol was consumed and period of consumption were analyzed. Three additional variables were analyzed for patients 18 years of age,
Results Participation in the three-stages of the survey is displayed in Fig. 2.
Table 2 Adjusted prevalence of active convulsive epilepsy by village of residence, Agincourt sub-district, 2008. Villagea
Cases of ACE
Adjusted prevalence
95% Confidence interval
A B C D E F G H I J K L M N O P Q R S T U V W X Y Overall
7 3 20 22 6 11 9 13 1 2 11 3 6 1 6 10 7 17 18 11 16 10 4 10 21 245
7.17 2.77 6.21 10.92 6.31 5.46 7.24 4.67 6.84 5.1 7.39 3.02 3.02 6.77 11.41 8.45 10.04 14.98 6.64 3.03 9.91 7.2 7.07 9.96 1.77 7.01
5.36—9.40 1.26—5.25 4.30—8.67 7.78—14.90 3.74—9.96 3.34—8.42 4.22—11.57 2.99—6.84 4.68—9.65 3.27—7.58 5.49—9.73 1.12—6.21 1.21—6.21 4.30—10.14 7.07—17.38 5.53—12.35 7.25—13.55 10.06—21.45 2.67—13.62 0.06—8.83 5.67—16.04 5.24—9.66 4.05—11.46 6.17—15.18 0.21—6.28 6.23—7.78
a Villages have been anonymized due to small village size and confidentiality of individuals with ACE.
786
R.G. Wagner et al. Eligible Population (2008 Agincourt Census) n=83,121
Stage 1
Deceased; n=45 Not found; n=258 Screened in Stage 1 n=82,818 (99.64%)
Positive; n= 546 (0.66%)
Population Sample n= 4500
Unwilling; n=4 Not found; n=27 Stage 2 Screened in Stage 2 n= 515 (94.23%)
Found; n= 3889 (91.44%)
Positive; n= 354 (68.74%)
Positive; n= 67 (1.72%) (38 also found positive in Stage 1)
Matched Controls n= 262 (90.34%)
Unwilling; n=4 Not found; n=27 Stage 3
Self-referred; found positive at various Stages
Assessed in Stage 3 n= 328 (92.66%)
Assessed in Stage 3 n= 58 (86.57%)
Positive (ACE) n= 245 (74.70%)
Positive (ACE) n= 26 (44.83%)
Assessed; n= 262 (100%)
Positive (ACE) n= 56 (33 screen negative in Stage 1)
(16 screened positive in Stage 1)
Figure 2
Study design schema and numbers of individuals at each stage, 2008.
Prevalence We identified 245 cases of ACE in the three-stage study. The unadjusted prevalence of ACE was 3.0/1,000 individuals (95%CI: 2.6—3.3) within the three-stage study. Adjusting for attrition and the sensitivity of the three-stage methodology (48.6%), the adjusted crude prevalence was 7.0/1,000 (95%CI: 6.4—7.6). The sensitivity of 48.6% for the threestage methodology was derived from a validation study performed in Kenya using clinical assessments as the gold standard (Ngugi et al., 2012). The prevalence, standardized to the age distribution of the US population in 2000 and adjusted for attrition and sensitivity of the three-stage method, was 8.1/1,000 individuals (95%CI: 7.5—8.7). Within the population sample of 3,889 (611 not found due to permanent out-migration from the study area) the crude prevalence was 6.7/1,000 individuals (95%CI: 4.1—9.3), while the adjusted prevalence (adjusting for attrition and sensitivity (76.7%) of the two-stage method against the gold standard (Ngugi et al., 2012)) was 9.8/1,000 individuals (95% CI: 6.9—13.4). The adjusted prevalence of ACE by village ranged from 1.8—15.0/1,000 individuals (Table 2) and this variation was statistically significant (p = 0.05), although the response rate for Stage 1 and Stage 2 did not vary between villages.
Age of onset of epilepsy 239 individuals reported age of onset of their seizures. The distribution of age of onset was left-skewed with 7.1% (n = 17) experiencing seizures within the first year of life (Fig. 3) and 30.8% (n = 74) before 5 years. The median age at onset was 12 years interquartile range was 0—23 years.
Factors associated with epilepsy Of the 311 cases of ACE identified (245 from the three-stage survey, 10 additional cases from the population sample, and
56 cases referred from other sources), 292 cases (94%) were matched to 260 controls. The matching was not 1:1 due to a high proportion of refusals in the controls.
Univariate analysis Within the univariate analysis, male sex, family history of seizures, a sibling with seizures, problems after delivery, loss of consciousness during head injury and history of snoring at least three nights per week were associated with ACE. ACE was not associated with exposure to any of the parasites tested and none of the cases had antibodies to T. solium. For individuals >18 years, neither alcohol consumption nor length of alcohol consumption was associated with ACE. For individuals
journal homepage: www.elsevier.com/locate/epilepsyres
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Ryan G. Wagner a,b,c,∗, Anthony K. Ngugi a,d,e, Rhian Twine a, Christian Bottomley f,g, Gathoni Kamuyu d, F. Xavier Gómez-Olivé b, Myles D. Connor b,h, Mark A. Collinson b,c, Kathleen Kahn a,b,c, Stephen Tollman a,b,c, Charles R. Newton a,b,d,i,j,k a
Studies of Epidemiology of Epilepsy in Demographic Surveillance Systems (SEEDS)—INDEPTH Network, Accra, Ghana b MRC/Wits Rural Public Health & Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa c Division of Epidemiology and Global Health, Department of Public Health and Clinical Medicine, Umeå University, Umeå, Sweden d KEMRI/Wellcome Trust Research Programme, Centre for Geographic Medicine Research—Coast, Kilifi, Kenya e Research Support Unit, Faculty of Health Sciences, Aga Khan University- East Africa, Nairobi, Kenya f Department of Infectious Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom g MRC Tropical Epidemiology Group, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, United Kingdom h Borders General Hospital, Melrose, United Kingdom i Neurosciences Unit, UCL Institute of Child Health, London, United Kingdom j Clinical Research Unit, London School of Hygiene and Tropical Medicine, London, United Kingdom k Department of Psychiatry, University of Oxford, Oxford, United Kingdom Received 26 May 2013; received in revised form 21 November 2013; accepted 14 January 2014 Available online 29 January 2014
KEYWORDS Epilepsy; Prevalence;
Summary Rationale: Epilepsy is among the most common neurological disorders worldwide. However, there are few large, population-based studies of the prevalence and risk factors for epilepsy in southern Africa.
∗ Corresponding author at: MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), PO Box 2 Acornhoek 1360 South Africa. Tel.: +27 13 795 5076; fax: +27 13 795 5076. E-mail address: [email protected] (R.G. Wagner).
0920-1211/$ — see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2014.01.004
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Case-control; Risk factors; Population-based
783
Methods: From August 2008 to February 2009, as part of a multi-site study, we undertook a threestage, population-based study, embedded within the Agincourt health and socio-demographic surveillance system, to estimate the prevalence and identify risk factors of active convulsive epilepsy (ACE) in a rural South African population. Results: The crude prevalence of ACE, after adjusting for non-response and the sensitivity of the screening method, was 7.0/1,000 individuals (95%CI 6.4—7.6) with significant geographic heterogeneity across the study area. Being male (OR = 2.3; 95%CI 1.6—3.2), family history of seizures (OR = 4.0; 95%CI 2.0—8.1), a sibling with seizures (OR = 7.0; 95%CI 1.6—31.7), problems after delivery (OR = 5.9; 95%CI 1.2—24.6), and history of snoring (OR = 6.5; 95%CI 4.5—9.5) were significantly associated with ACE. For children, their mother’s exposure to some formal schooling was protective (OR = 0.30; 95%CI 0.11—0.84) after controlling for age and sex. Human immunodeficiency virus was not found to be associated with ACE. Conclusions: ACE is less frequent in this part of rural South Africa than other parts of sub-Saharan Africa. Improving obstetric services could prevent epilepsy. The relationship between snoring and ACE requires further investigation, as does the relative contribution of genetic and environmental factors to examine the increased risk in those with a family history of epilepsy. © 2014 Elsevier B.V. All rights reserved.
Introduction Epilepsy is one of the most common neurological disorders in the world, affecting about 69 million people worldwide, with 90 percent living in low- and middle-income countries (LMICs) (Ngugi et al., 2010). It contributes nearly one percent to the global burden of disease (Murray et al., 2012), and 20 percent of the global burden of epilepsy is in Africa (World Health Organization, 2004). While these figures suggest a large burden of epilepsy in Africa, they are derived from a limited number of studies that employ different case definitions and methodologies. Studies suggest that utilizing hospital records in LMICs to detect epilepsy under-estimates the prevalence by at least 80 percent due to limited health care utilization by people with epilepsy (PWE) in these settings (Osuntokun et al., 1987). Thus, population-based surveys are frequently used to estimate prevalence, though not without limitations, including the absence of well-demarcated populations and vital statistics registries. This limitation, coupled with the lack of trained medical personnel available to make the diagnosis of epilepsy, makes estimating the burden of epilepsy in sub-Saharan Africa unusually challenging. A recent systematic review and meta-analysis highlighted the significant variation in the prevalence of epilepsy between high-income countries and LMICs, with a higher prevalence in LMICs, especially in rural settings (Ngugi et al., 2010). The authors suggest that study size and the economic development level of the study country largely explain the heterogeneity, although increases in obstetric injury, head injuries, and infections and infestations of the central nervous system (CNS) (Newton & Garcia, 2012), such as toxoplasmosis and toxocara (Wagner & Newton, 2009), malaria (Carter et al., 2004) and human immunodeficiency virus (HIV) are thought to contribute (Ngugi et al., 2010), but there is little data from South Africa where the prevalence of HIV is very high. As part of a multi-centre study on the epidemiology of epilepsy in demographic sites (SEEDS) (Ngugi et al., 2013), we conducted a three-stage, population-based survey and a case-control study to determine the prevalence of and risk factors for active convulsive epilepsy (ACE) in rural South
Africa. In particular, we were interested in examining the risk factors for ACE in a non-malaria endemic area, particularly HIV since it has a high prevalence in South Africa.
Methodology Population and study area The study was conducted in the rural Agincourt health sub-district, in which the Agincourt health and sociodemographic surveillance system (HDSS) operates and is located 500 kilometers northeast of Johannesburg (Fig. 1). The Agincourt HDSS was established in 1992 as a research platform to inform health and development policy through evidence-based research (Kahn et al., 2012). The population has been enumerated through an annual census update, following baseline measurement in 1992 and captures vital statistics including births, deaths, and in- and out-migrations. In 2008, the population was 83,121 individuals in 15,841 households and 25 villages on 420 km2 of semi-arid scrubland. The site forms part of a former bantustan, or ethnic ‘homeland’ implemented during the Apartheid era. The population is mainly Xi-Tsonga speaking, with one-third (32.8%) originally from Mozambique. The leading causes of death ascertained through the Agincourt HDSS are HIV/AIDS, cardiovascular disease and trauma (road traffic accidents, assaults) (Tollman et al., 2008). Six government clinics, one larger government health centre, and one public-private community health centre, with its main focus being HIV and tuberculosis, provide primary health care services for the population. Referrals are to three government district hospitals located 25 to 55 km from the sub-district.
Ethics Ethical clearance for the study was received from the Human Research Ethics Committee of the University of the Witwatersrand, Johannesburg, South Africa (Clearance number: M080455) and the Mpumalanga Province Department of
784
R.G. Wagner et al.
Figure 1 Map of South Africa, Bushbuckridge Municipality, and the Agincourt sub-district and health and socio-demographic surveillance system, 2008.
Health’s Research and Ethics committee. Written informed consent was sought from each participant in the study. Parental or guardian informed consent was sought in the case of children or patients with cognitive impairment, with verbal assent being sought from children and cognitively impaired patients.
Case definition of active convulsive epilepsy (ACE) We identified patients with Active convulsive epilepsy (ACE) since convulsive epilepsies are associated with increased morbidity, mortality (Diop et al., 2005) and greater social stigma than non-convulsive epilepsies (World Health Organization, 2005). ACE was defined as having two or more unprovoked convulsive seizures occurring more than 24 hours apart, with at least one seizure occurring in the 12 months preceding the study or currently taking antiepileptic drugs (AEDs) (Edwards et al., 2008).
and study neurologist (CRN) on the administration of this questionnaire. Individuals who had experienced at least two seizures in their lifetime and at least one seizure with abnormal movement (aimed at identifying convulsive epilepsy) in the preceding 12 months or currently taking AEDs were invited for further assessment in the epilepsy clinic within the following week (Stage III). In Stage III (September 2008—February 2009), sociodemographic information was obtained and a specially trained nurse made the diagnosis of ACE, based on a positive clinical history provided by the patient’s caregiver or family member. A clinical examination was completed to identify neurological abnormalities and physical co-morbidities. Blood was drawn from each patient for serological testing. A neurologist (CRN) confirmed the diagnosis of epilepsy by reviewing all of the patient records and examining selected patients.
Population sample Procedures Between August and November 2008, two questions (‘Do you have fits or has someone ever told you that you have fits?’ and ‘Do you experience episodes in which you legs or arms have jerking movements or fall to the ground and lose consciousness?’), used previously in a similar study (Edwards et al., 2008), were administered by census fieldworkers to a senior member of each household (n = 15,841) on behalf of each member of the household. The questions had been piloted previously in the local language (Xi-Tsonga) at a district hospital and had a high specificity (100%) and sensitivity (98%). The questions were included in the 2008 annual Agincourt HDSS census update round. The two questions within the census update constituted Stage I of the study and sought to identify all people who had experienced a seizure prior to 1 August 2008 (the prevalence day). In Stage II (August—December 2008), trained fieldworkers visited all individuals who responded ‘yes’ to either of the two questions in Stage I and asked a detailed questionnaire, based on previous studies (Placencia et al., 1992), about the characteristics and history of the seizures. Fieldworkers were specifically trained by the study managers (RGW, AKN)
In addition to identifying cases of epilepsy through the population screen (Stage I), a random sample of 4,500 individuals (all ages and both sexes), from the Agincourt HDSS, were screened with the Stage II tool. This population sample simulated the traditional two-stage epilepsy prevalence study design (Placencia et al., 1992) and was used to evaluate the three-stage methodology of this study. Those who screened positive were referred for assessment in Stage III and included as cases in the case-control study. Furthermore, individuals who were referred to the epilepsy clinic by sources (clinic staff and community leaders) other than the three-stage study were clinically assessed and those found to have ACE were included as cases in the case-control study.
Selection of controls for case control study Controls were selected randomly from those that screened negative for epilepsy in the population sample and frequency matched to cases diagnosed with ACE during Stage III of the study (on age-bands 0—5, 6—12, 13—18, 19—28, 29—49, and 50+ years). If the selected control was not found
Prevalence and risk factors for active convulsive epilepsy in rural northeast South Africa Table 1
785
Crude and adjusted prevalence of ACE by age and sex, Agincourt sub-district, South Africa 2008.
Age band
Cases of ACE
Crude prevalence
95% Confidence interval
Adjusted prevalence
95% Confidence interval
0 to 5 years 6 to 12 years 13 to 18 years 19 to 28 years 29 to 49 years 50+ years Male Female
10 29 34 47 87 38 129 116
0.87 2.23 2.76 2.65 4.70 3.94 3.24 2.69
0.33—1.41 1.42—3.03 1.83—3.68 1.89—3.40 3.71—5.69 2.68—5.19 2.68—3.80 2.20—3.18
2.18 4.79 6.38 6.23 10.95 9.55 7.43 6.50
0.88—3.48 3.09—6.52 4.32—8.42 4.47—7.98 8.74—13.13 6.69—12.41 5.53—7.82 4.53—6.54
after three attempted visits, another control was selected randomly as a replacement. For each control selected, a clinical history and blood sample were taken and a clinical examination performed.
variables on whether alcohol was consumed and period of consumption were analyzed. Three additional variables were analyzed for patients 18 years of age,
Results Participation in the three-stages of the survey is displayed in Fig. 2.
Table 2 Adjusted prevalence of active convulsive epilepsy by village of residence, Agincourt sub-district, 2008. Villagea
Cases of ACE
Adjusted prevalence
95% Confidence interval
A B C D E F G H I J K L M N O P Q R S T U V W X Y Overall
7 3 20 22 6 11 9 13 1 2 11 3 6 1 6 10 7 17 18 11 16 10 4 10 21 245
7.17 2.77 6.21 10.92 6.31 5.46 7.24 4.67 6.84 5.1 7.39 3.02 3.02 6.77 11.41 8.45 10.04 14.98 6.64 3.03 9.91 7.2 7.07 9.96 1.77 7.01
5.36—9.40 1.26—5.25 4.30—8.67 7.78—14.90 3.74—9.96 3.34—8.42 4.22—11.57 2.99—6.84 4.68—9.65 3.27—7.58 5.49—9.73 1.12—6.21 1.21—6.21 4.30—10.14 7.07—17.38 5.53—12.35 7.25—13.55 10.06—21.45 2.67—13.62 0.06—8.83 5.67—16.04 5.24—9.66 4.05—11.46 6.17—15.18 0.21—6.28 6.23—7.78
a Villages have been anonymized due to small village size and confidentiality of individuals with ACE.
786
R.G. Wagner et al. Eligible Population (2008 Agincourt Census) n=83,121
Stage 1
Deceased; n=45 Not found; n=258 Screened in Stage 1 n=82,818 (99.64%)
Positive; n= 546 (0.66%)
Population Sample n= 4500
Unwilling; n=4 Not found; n=27 Stage 2 Screened in Stage 2 n= 515 (94.23%)
Found; n= 3889 (91.44%)
Positive; n= 354 (68.74%)
Positive; n= 67 (1.72%) (38 also found positive in Stage 1)
Matched Controls n= 262 (90.34%)
Unwilling; n=4 Not found; n=27 Stage 3
Self-referred; found positive at various Stages
Assessed in Stage 3 n= 328 (92.66%)
Assessed in Stage 3 n= 58 (86.57%)
Positive (ACE) n= 245 (74.70%)
Positive (ACE) n= 26 (44.83%)
Assessed; n= 262 (100%)
Positive (ACE) n= 56 (33 screen negative in Stage 1)
(16 screened positive in Stage 1)
Figure 2
Study design schema and numbers of individuals at each stage, 2008.
Prevalence We identified 245 cases of ACE in the three-stage study. The unadjusted prevalence of ACE was 3.0/1,000 individuals (95%CI: 2.6—3.3) within the three-stage study. Adjusting for attrition and the sensitivity of the three-stage methodology (48.6%), the adjusted crude prevalence was 7.0/1,000 (95%CI: 6.4—7.6). The sensitivity of 48.6% for the threestage methodology was derived from a validation study performed in Kenya using clinical assessments as the gold standard (Ngugi et al., 2012). The prevalence, standardized to the age distribution of the US population in 2000 and adjusted for attrition and sensitivity of the three-stage method, was 8.1/1,000 individuals (95%CI: 7.5—8.7). Within the population sample of 3,889 (611 not found due to permanent out-migration from the study area) the crude prevalence was 6.7/1,000 individuals (95%CI: 4.1—9.3), while the adjusted prevalence (adjusting for attrition and sensitivity (76.7%) of the two-stage method against the gold standard (Ngugi et al., 2012)) was 9.8/1,000 individuals (95% CI: 6.9—13.4). The adjusted prevalence of ACE by village ranged from 1.8—15.0/1,000 individuals (Table 2) and this variation was statistically significant (p = 0.05), although the response rate for Stage 1 and Stage 2 did not vary between villages.
Age of onset of epilepsy 239 individuals reported age of onset of their seizures. The distribution of age of onset was left-skewed with 7.1% (n = 17) experiencing seizures within the first year of life (Fig. 3) and 30.8% (n = 74) before 5 years. The median age at onset was 12 years interquartile range was 0—23 years.
Factors associated with epilepsy Of the 311 cases of ACE identified (245 from the three-stage survey, 10 additional cases from the population sample, and
56 cases referred from other sources), 292 cases (94%) were matched to 260 controls. The matching was not 1:1 due to a high proportion of refusals in the controls.
Univariate analysis Within the univariate analysis, male sex, family history of seizures, a sibling with seizures, problems after delivery, loss of consciousness during head injury and history of snoring at least three nights per week were associated with ACE. ACE was not associated with exposure to any of the parasites tested and none of the cases had antibodies to T. solium. For individuals >18 years, neither alcohol consumption nor length of alcohol consumption was associated with ACE. For individuals
